Text Mining for Patterns and Associations on Functions, Relationships and Prioritization in Services Reflected in National Health Insurance Programs

Document Type

Conference Proceeding

Publication Date



The new Philippine Universal Health Care (UHC) Act was adopted in February 2019 and was slated for implementation in January 2020. However, the ongoing pandemic has thrown the national health insurance program under close scrutiny as the single biggest source of funding for hospitalization or personal health care. An inductive study using text analysis is utilized to discover implementation patterns over time of the national health insurance program’s (NHIP) policies extracted as functions and relationships in publicly issued executive circulars. Standard text mining methods were used to extract publicly available PhilHealth Circulars from 2000–2020. Principal component analysis was used to determine clusters per time period, where time period covered years served each administration from 2000–2020. Result of PCA was used to determine functions. Within each cluster, frequent pattern algorithm (FP growth) was used to determine words that are most frequently associated with each other. This methodology was used to determine policy focus on the following services: membership, benefits, contributions, collection, and accreditation. Results indicate that the first four periods prioritized on accreditation while the remaining three periods focused on membership and benefits. The results also suggest insights on the use of text mining on policy-settings, on change trajectories, if any, and implications for the implementation of the new UHC Act.